Anxiety disorders are common and, if untreated, influence children's lives. In randomized clinical trials (RCT), research has shown that cognitive behavioral therapy (CBT) benefits anxiety-disordered youth. We are currently comparing individual CBT, family CBT, and an education/support/attention condition. Though there has been an increase in outcome research, there has been very little study of the treatment process. With audio and video taped sessions from 154 cases that have received individual CBT in previous RCTs and 50 cases from the current trial, we will examine the relations between process variables and treatment outcome within CBT. The process variables are (1) therapeutic alliance, (2) child involvement and (3) therapist flexibility with the manual-based treatment. These three variables will be measured from sessions 1-4 (earlier process) and sessions 5-8 (later process) for the entire data set. In addition, the three process variables will be measured from sessions 9-16 for the 50 cases in the current RCT. Raters will rate the entire sessions a third at a time from a total of 1600 sessions of individual CBT using the Child Psychotherapy Process Scale, Child Involvement rating scales and the Modified Protocol Adherence Checklist. It is hypothesized that a strong therapeutic alliance and high levels of child involvement will predict positive treatment outcome. The identification of mediating influences (e.g., later involvement, therapist flexibility) will further enable the outcome researcher and the practicing clinician to develop a better understanding of the causal mechanisms between therapy and the accrual of beneficial outcomes. To evaluate potential process variables that contribute to patient attrition, we will compare the early process variables for children completing treatment and those who discontinue. It is hypothesized that non-completers will have the experienced lower levels of alliance, flexibility, and involvement.